Support Vector Regression to Accelerate Design and Crosspolar Optimization of Shaped-Beam Reflectarray Antennas for Space Applications

Abstract

A machine learning technique is applied to the design and optimization of reflectarray antennas to considerably accelerate computing time without compromising accuracy. In particular, Support Vector Machines (SVMs), automatic learning structures that are able to deal with regression problems, are employed to obtain surrogate models of the reflectarray element to substitute the full-wave analysis tool for the characterization of the unit cell in the design and optimization algorithms. The analysis, design and optimization of a very large reflectarray antenna for Direct Broadcast Satellite applications are accelerated up to three orders of magnitude. This is here demonstrated with three examples: one showing the design of a reflectarray; and two for the crosspolar optimization, one with one coverage for each linear polarization (Europe and the Middle East) and another with a Middle East coverage working in dual-linear polarization. The accuracy of the proposed approach is validated by means of a comparison of the final designs with full-wave simulations based on local periodicity obtaining good agreement. The crosspolar dicrimination and crosspolar isolation are greatly improved using the SVMs while considerably reducing computing time.

A review on the design and optimization of antennas using machine learning algorithms and techniques

“…A surrogate model was obtained using the SVMs. SVM was used in References 172‐174 to design shaped‐beam reflectarrays and for modeling dual‐polarized reflectarray unit cells in Reference 175. Using SVM, the computational burden resulting from the use of Full‐Wave Local‐Periodicity for the design and analysis was reduced.…”

Simulation-Driven Antenna Modeling by Means of Response Features and Confined Domains of Reduced Dimensionality

“…Miscellaneous frameworks have been developed including strictly algorithmic methods (e.g., based on selective suppression of sensitivity updates through finite differentiation [28], [29]) or the employment of adjoint sensitivities [30]. An alternative approach is to exploit surrogate models (or metamodels) [31]- [41].…”

“…The primary reasons for their widespread use are the following: (i) their construction requires no physical insight; (ii) they are easily transferable between application areas, and (iii) they are readily accessible (e.g., SUMO [42], DACE [43], UQlab [44]). Plenitude of the data-driven surrogate modelling techniques have been developed, e.g., kriging [35], radial basis functions (RBF) [36], neural networks [37]- [39], Gaussian process [40] or support vector regression [41]. A primary limiting factor of data-driven models is the curse of dimensionality.…”

Variable‐fidelity modeling of antenna input characteristics using domain confinement and two‐stage Gaussian process regression surrogates

“…Some widely used techniques include polynomial regression, 26 kriging, 27 radial basis function interpolation, 28 Gaussian process regression (GPR), 29 neural networks, 30 polynomial chaos expansion, 31 and support vector regression. 32 All of these methods are affected by the curse of dimensionality, that is, a rapid growth of the number of training data samples required to render a usable model as a function of the number of system parameters and their ranges. In the case of antennas, typically characterized by highly nonlinear responses, conventional surrogates can be constructed for structures described by a few parameters and within narrow ranges thereof.…”